Kinetics and Mechanism of •NO2 Reacting with Various Oxidation

Published on Web 11/11/2004

Kinetics and Mechanism of •NO2 Reacting with Various Oxidation States of Myoglobin Sara Goldstein,*,† Gabor Merenyi,§ and Amram Samuni‡ Contribution from the Department of Physical Chemistry, The Hebrew UniVersity of Jerusalem, Jerusalem 91904, Department of Molecular Biology, The Hebrew UniVersity of Jerusalem-Hadassah Medical School, Jerusalem 91120, Israel, and Department of Chemistry, Nuclear Chemistry, The Royal Institute of Technology, S-10044 Stockholm 70, Sweden Received June 28, 2004; E-mail: [email protected]

Abstract: Nitrogen dioxide (•NO2) participates in a variety of biological reactions. Of great interest are the reactions of •NO2 with oxymyoglobin and oxyhemoglobin, which are the predominant hemeproteins in biological systems. Although these reactions occur rapidly during the nitrite-catalyzed autoxidation of hemeproteins, their roles in systems producing •NO2 in the presence of these hemeproteins have been greatly underestimated. In the present study, we employed pulse radiolysis to study directly the kinetics and mechanism of the reaction of oxymyoglobin (MbFeIIO2) with •NO2. The rate constant of this reaction was determined to be (4.5 ( 0.3) × 107 M-1s-1, and is among the highest rate constants measured for •NO with any biomolecule at pH 7.4. The interconversion among the various oxidation states of myoglobin 2 that is prompted by nitrogen oxide species is remarkable. The reaction of MbFeIIO2 with •NO2 forms MbFeIIIOONO2, which undergoes rapid heterolysis along the O-O bond to yield MbFeVdO and NO3-. The perferryl-myoglobin (MbFeVdO) transforms rapidly into the ferryl species that has a radical site on the globin (•MbFeIVdO). The latter oxidizes another oxymyoglobin (104 M-1s-1 < k17 < 107 M-1s-1) and generates equal amounts of ferrylmyoglobin and metmyoglobin. At much longer times, the ferrylmyoglobin disappears through a relatively slow comproportionation with oxymyoglobin (k18 ) 21.3 ( 5.3 M-1s-1). Eventually, each •NO2 radical converts three oxymyoglobin molecules into metmyoglobin. The same intermediate, namely MbFeIIIOONO2, is also formed via the reaction peroxynitrate (O2NOO-/O2NOOH) with metmyoglobin (k19 ) (4.6 ( 0.3) × 104 M-1s-1). The reaction of •NO2 with ferrylmyoglobin (k20 ) (1.2 ( 0.2) × 107 M-1s-1) yields MbFeIIIONO2, which in turn dissociates (k21 ) 190 ( 20 s-1) into metmyoglobin and NO3-. This rate constant was found to be the same as that measured for the decay of the intermediate formed in the reaction of MbFeIIO2 with •NO, which suggests that MbFeIIIONO2 is the intermediate observed in both processes. This conclusion is supported by thermokinetic arguments. The present results suggest that hemeproteins may detoxify •NO2 and thus preempt deleterious processes, such as nitration of proteins. Such a possibility is substantiated by the observation that the reactions of •NO2 with the various oxidation states of myoglobin lead to the formation of metmyoglobin, which, though not functional in the gas transport, is nevertheless nontoxic at physiological pH.

Introduction

Nitrogen dioxide (•NO2) participates in a variety of biological reactions. It is formed through the decomposition of peroxynitrite in the absence and presence of CO2,1,2 during autoxidation of •NO,3-5 oxidation of nitrite by hemeproteins,6-16 and in the †

Department of Physical Chemistry, The Hebrew University of Jerusa-

lem. ‡ Department of Molecular Biology, The Hebrew University of Jerusalem-Hadassah Medical School. § Department of Chemistry, Nuclear Chemistry, The Royal Institute of Technology.

(1) Goldstein, S.; Czapski, G. J. Am. Chem. Soc. 1998, 120, 3458-3463. (2) Merenyi, G.; Lind, J.; Goldstein, S.; Czapski, G. J. Phys. Chem. A 1999, 103, 5685-5691. (3) Ford, P. C.; Wink, D. A.; Stanbury, D. M. FEBS Lett. 1993, 326, 1-3. (4) Lewis, R. S.; Tannenbaum, S. R.; Deen, W. M. J. Am. Chem. Soc. 1995, 117, 3933-3939. (5) Goldstein, S.; Czapski, G. J. Am. Chem. Soc. 1995, 117, 12078-12084. (6) Arduini, A.; Mancinelli, G.; Radatti, G. L.; Hochstein, P.; Cadenas, E. Arch. Biochem. Biophys. 1992, 294, 398-402. (7) Wade, R. S.; Castro, C. E. Chem. Res. Toxicol. 1996, 9, 1382-1390. 15694

9

J. AM. CHEM. SOC. 2004, 126, 15694-15701

reaction of peroxynitrite with metalloproteins15,17-21 and particularly hemeproteins.11,22 Nitrogen dioxide is produced through (8) Byun, J.; Mueller, D. M.; Fabjan, J. S.; Heinecke, J. W. FEBS Lett. 1999, 455, 243-246. (9) Reszka, K. J.; Matuszak, Z.; Chignell, C. F.; Dillon, J. Free Radic. Biol. Med. 1999, 26, 669-678. (10) Burner, U.; Furtmuller, P. G.; Kettle, A. J.; Koppenol, W. H.; Obinger, C. J. Biol. Chem. 2000, 275, 20597-20601. (11) Bourassa, J. L.; Ives, E. P.; Marqueling, A. L.; Shimanovich, R.; Groves, J. T. J. Am. Chem. Soc. 2001, 123, 5142-5143. (12) Lissi, E. Free Radic. Biol. Med. 1998, 24, 1535-1536. (13) Pietraforte, D.; Salzano, A. M.; Marino, G.; Minetti, M. Amino Acids 2003, 25, 341-350. (14) Herold, S. Free Radic. Biol. Med. 2004, 36, 565-579. (15) Nicolis, S.; Monzani, E.; Roncone, R.; Gianelli, L.; Casella, L. Chem. Eur. J. 2004, 10, 2281-2290. (16) Monzani, E.; Roncone, R.; Galliano, M.; Koppenol, W. H.; Casella, L. Eur. J. Biochem. 2004, 271, 895-906. (17) Lee, J. B.; Hunt, J. A.; Groves, J. T. J. Am. Chem. Soc. 1998, 120, 74937501. (18) Ferrer-Sueta, G.; Batinic-Haberle, I.; Spasojevic, I.; Fridovich, I.; Radi, R. Chem. Res. Toxicol. 1999, 12, 442-449. (19) Groves, J. T.; Lee, J.; Hunt, J. A.; Shimanovich, R.; Jin, N. J. Inorg. Biochem. 1999, 74, 28-28. 10.1021/ja046186+ CCC: $27.50 © 2004 American Chemical Society

Kinetics and Mechanism of •NO2

the reaction of nitrite with the ferryl oxidation state of proteins or porphyrins15,23,24 as well as through the reaction of peroxynitrite with the ferric oxidation state of these biomolecules11,17,22 and with oxyhemoglobin.25 Of great interest are the reactions of •NO2 with hemeproteins and particularly with oxymyoglobin and oxyhemoglobin, which are the predominant hemeproteins in biological systems. Although these reactions occur rapidly during the nitrite-catalyzed autoxidation of hemeproteins,12,26-29 they were nevertheless underestimated in systems where these hemeproteins reacted with peroxynitrite in the absence and presence of CO2.25,30-32 An attempt to determine the rate constant for the reaction of •NO with oxymyoglobin by injection of •NO gas into oxy2 2 myoglobin solutions failed because of the rapid hydrolysis of •NO to yield NO - and NO -.7 Furthermore, there is no 2 2 3 agreement on the mechanism of this reaction. Kosaka et al.26,27 have suggested that the reaction of oxyhemoglobin with •NO2 forms methemoglobin and NO2-, whereas others7,12,28,29 have proposed that the reaction forms perferryl-myoglobin and NO3-. Nitration of proteins is currently a subject of great interest, since its occurrence has been detected in various pathological conditions.33-36 It has been demonstrated that the formation of •NO2 in the reaction of nitrite with the ferryl oxidation state of proteins can lead to nitration of tyrosine residues in proteins.11,13-15,23,25,37-39 However, if oxymyoglobin and oxyhemoglobin react rapidly with •NO2, they may compete with the tyrosyl radical for •NO2 and inhibit tyrosine nitration. Nevertheless, while they can serve as anti-nitrosative agents, they might also act as prooxidants if other toxic species are produced during the reaction of •NO2 with these hemeproteins. In the present study, we employed pulse radiolysis to study directly the kinetics and mechanism of the reaction of oxymyoglobin with •NO2. We found that the rate constant for this reaction is among the fastest known for reactions of •NO2 with any biomolecule under physiological conditions. The biological implications are also discussed. (20) Shimanovich, R.; Hannah, S.; Lynch, V.; Gerasimchuk, N.; Mody, T. D.; Magda, D.; Sessler, J.; Groves, J. T. J. Am. Chem. Soc. 2001, 123, 36133614. (21) Ferrer-Sueta, G.; Vitturi, D.; Batinic-Haberle, I.; Fridovich, I.; Goldstein, S.; Czapski, G.; Radi, R. J. Biol. Chem. 2003, 278, 27432-27438. (22) Herold, S.; Shivashankar, K. Biochemistry 2003, 42, 14036-14046. (23) Kilinc, K.; Kilinc, A.; Wolf, R. E.; Grisham, M. B. Biochem. Biophys. Res. Commun. 2001, 285, 273-276. (24) Herold, S.; Rehmann, F. J. K. Free Radic. Biol. Med. 2003, 34, 531-545. (25) Romero, N.; Radi, R.; Linares, E.; Augusto, O.; Detweiler, C. D.; Mason, R. P.; Denicola, A. J. Biol. Chem. 2003, 278, 44049-44057. (26) Kosaka, H.; Imaizumi, K.; Tyuma, I.Biochim. Biophys. Acta 1982, 702, 237-241. (27) Kosaka, H.; Tyuma, I. EnViron. Health Perspect. 1987, 73, 147-151. (28) Doyle, M. P.; Herman, J. G.; Duykstra, R. L. J. Free Rad. Biol. Med. 1988, 1, 145-153. (29) Steinhaus, R. K.; Baskin, S. I.; Clark, J. H.; Kirby, S. D. J. Appl. Toxicol. 1990, 10, 345-351. (30) Minetti, M.; Scorza, G.; Pietraforte, D. Biochemistry 1999, 38, 2078-2087. (31) Exner, M.; Herold, S. Chem. Res. Toxicol. 2000, 13, 287-293. (32) Herold, S.; Exner, M.; Boccini, F. J. Am. Chem. Soc. 2003, 16, 390-402. (33) MacMillan-Crow, L. A.; Crow, J. P.; Kerby, J. D.; Beckman, J. S.; Thompson, J. A. Proc. Natl. Aca. Sci. USA 1996, 93, 11853-11858. (34) Ischiropoulos, H. Arch. Biochem. Biophys. 1998, 356, 1-11. (35) Bachschmid, M.; Thurau, S.; Zou, M.-H.; Ulltich, V. FASEB J. 2003, 17, 914-916. (36) Shishehbor, M. H.; Aviles, R. J.; Brennan, M.-L.; Fu, X.; Goormastic, M.; Pearce, G. L.; Gokce, N.; Keaney, J. F., Jr; Penn, M. S.; Sprecher, D. L.; Vita, J. A.; Hazen, S. L. JAMA 2003, 289, 1675-1680. (37) Brennan, M.-L.; Wu, W.; Fu, X.; Shen, Z.; Song, W.; Frost, H.; Vadseth, C.; Narine, L.; Lenkiewicz, E.; Borchers, M. T.; Lusis, A. J.; Lee, J. J.; Lee, N. A.; Abu-Soud, H. M.; Ischiropoulos, H.; Hazen, S. L. J. Biol. Chem. 2002, 277, 17415-17427. (38) Bian, K.; Gao, Z.; Weisbrodt, N.; Murad, F. PNAS 2003, 100, 5712-5717. (39) Grzelak, A.; Balcerczyk, A.; Mateja, A.; Bartosz, G. Biochim. Biophys. Acta-Gen. Subj. 2001, 1528, 97-100.

ARTICLES

Experimental Section

Materials. All chemicals were of analytical grade and were used as received. Solutions were prepared with distilled water, which was further purified using a Milli-Q water purification system. Myoglobin from horse heart and reduced β-nicotinamide adenine dinucleotide (NADH) grade III from yeast were obtained from Sigma. The concentration of NADH was determined spectrophotometrically using 340 ) 6200 M-1cm-1. Catalase (2 mg/mL, about 130 000 U/mL) was obtained from Boehringer Mannheim. Isolated metmyoglobin (MbFeIIIOH2) or oxymyoglobin (MbFeIIO2) solutions were prepared by adding excess of ferricyanide or dithionite, respectively, to commercial myoglobin, which always contains some metmyoglobin as well. These solutions were then subjected to chromatographic separation through a Sephadex G-25 column using 5-50 mM phosphate buffer (PB) as an eluent. Ferrylmyoglobin (MbFeIVd O) was prepared by adding excess of H2O2 to metmyoglobin. The solution, which was stored on ice, proved stable for the duration of the experiment. Catalase was added immediately before each experiment to remove excess of H2O2. Some buffered solutions for radiation experiments were saturated with N2O (25 mM) or with a mixture of N2O and O2 prior to the addition of small volumes of the protein, the latter having been prepared at relatively high concentrations. Hence, under all experimental conditions the concentration of N2O was higher than 16 mM while that of O2 did not exceed 0.4 mM. Spectrophotometric Determinations. Spectral properties of the oxidation states of myoglobin were monitored by recording the UV-visible absorption with 1, 2 or 10-mm optical path lengths using a Hewlett-Packard 8453 UV-vis diode array spectrophotometer. The concentration of each redox state was obtained from its absorption at neutral pH. The concentration of MbFeIIO2 was determined spectrophotometrically at 410, 543, and 582 nm using  ) 128, 13.9, and 14.4 mM-1cm-1, respectively.40 The concentration of MbFeIIIOH2 was calculated from its absorption at 406 and 503 nm using  ) 188 and 10.2 mM-1cm-1, respectively.40 The concentration of MbFeIVdO was determined using 421 ) 111 mM-1cm-1. Finally, the H2O2 content was derived from its absorption at 240 nm using  ) 39.4 M-1cm-1. Since the spectra of metmyoglobin and ferrylmyoglobin are pH-dependent, the difference spectra between these two redox states at each of the pH values studied were determined in order to assay the oxidation/reduction yields. In those cases where two or three redox states were involved, we calculated the concentration of each species using the absorptions of the mixture at 500 (or 582), 602 and 630 nm and the extinction coefficients of each redox state at these wavelengths. Methods. Pulse radiolysis experiments were carried out using a 5-MeV Varian 7715 linear accelerator (0.1-0.5 µs electron pulses, 200 mA current). A 200 W Xe lamp produced the analyzing light. Appropriate cutoff filters were used to minimize photochemistry. All measurements were made at room temperature using a 1-cm spectrosil cell and applying three light passes. The dose per pulse was determined with the Fricke dosimeter using G(Fe3+) ) 1.56 × 10-6 M Gy-1 and 302(Fe3+) ) 2197 M-1cm-1. Steady-state γ-irradiation was carried out at room temperature using a 137Cs source with a dose rate of 9.4 Gy min-1 as determined by the Fricke dosimetry. (40) Antonini, E.; Brunori, M. Hemoglobin and Myoglobin in Their Reactions with Ligands.; North-Holland: Amsterdam, 1971. J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004 15695

Goldstein et al.

ARTICLES

Generation of Nitrogen Dioxide. •NO2 was generated upon irradiation of buffered solutions containing 1 mM nitrite and saturated with a mixture of N2O and O2 at pH > 7. Under such conditions eaq- and •OH are converted into •NO2 via the following reactions γ

H2O 98 e-aq(2.6), •OH (2.7), H• (0.6), H2 (0.45), H2O2 (0.7), H3O+ (2.6) (1) The numbers in parentheses are G-values, which represent the concentrations of the species (in 10-7 M Gy-1), and are higher by about 7% in the presence of high solute concentrations.

Results and Discussion

eaq- + N2O + H2O f •OH + N2 + OH- k2 ) -1 -1 4

9.1 × 10 M s 9

(2)

•

OH + NO2- f •NO2 + OH- k3 ) 5.3 × 109 M-1s-1 41 (3)

radicals are converted into O2•- (reaction 4) and •NO (reaction 5), and the corresponding yields depend on the amount of O2 present in the solution.

•H

H• + O2 f H+ + O2•- k4 ) 1.2 × 1010 M-1s-1 41 (4) H• + NO2- f •NO + OH- k5 ) 1.5 × 109 M-1s-1 42 (5) In those cases where the tested compound, such as MbFeIIO2, reacts with nitrite, the •NO2 radical was formed upon irradiation of aerated solutions containing 20 mM NO 3-, 40 mM tert-butyl alcohol and 12 mM PB at pH 7.1-7.3. Under such conditions H• is converted into O2•- (reaction 4), •OH is converted into •OOCH C(CH ) OH (t-BuOO•) (reactions 6 and 7) 2 3 2 •

OH + (CH3)3COH f •CH2C(CH3)2OH + H2O k6 ) 6 × 108 M-1s-1 41 (6)

•

CH2C(CH3)2OH + O2 f •OOCH2C(CH3)2OH k7 ) 1.6 × 109 M-1s-1 41 (7)

and eaq- is converted into •NO2 through reactions 8 and 9.

eaq- + NO3- f NO32- k8 ) 9.7 × 109 M-1s-1 41

(8)

NO32- + H2PO4- f •NO2 + OH- + HPO42- k9 ) 5 × 108 M-1s-1 41 (9) However, in systems where tert-butyl alcohol is used to trap radicals, the side reactions of t-BuOO• (eqs 10-11) should be considered. The reaction of •NO2 with a tested compound has to compete with its reaction with t-BuOO• (reaction 10) and with O2•- (reaction 12), where the latter is also produced in reaction 11.

•OH

t-BuOO• + •NO2 h t-BuOONO2 k10 ) 7 × 108 M-1s-1 k-10 ≈ 0.07 s-1 43 (10) t-BuOO• + t-BuOO• f 0.27 O2•- + products 2k11 ) 8 × 108 M-1s-1 43 (11) •

NO2 + O2•- f O2NOO- k12 ) 4.5 × 109 M-1s-1 44 (12)

15696 J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004

In the absence of nitrate, and when the solutions are saturated with 80% N2O and 20% O2, both eaq- and •OH are converted into t-BuOO• and H• is converted into O2•-. Generation of Peroxynitrate (O2NOOH/O2NOO-). Peroxynitrate was generated upon pulse irradiation of aerated solutions containing 40 mM nitrate, 0.1 M formate and 2 mM acetate buffer at pH 4.8-5.0 or 24 mM PB at pH 5.9-7.2. Under such conditions, eaq- is converted into •NO2 (reactions 8, 9), •OH and H• react with HCO2- to form CO2•-, which reduces dioxygen to O2•-, and peroxynitrate is formed via reaction 12. Consequently, G(O2NOOH) + G(O2NOO-) ) G(•NO2), where pKa(O2NOOH) ) 5.9.45 Oxymyoglobin and metmyoglobin are inert toward NO3but the former is readily oxidized by nitrite. Therefore, the reaction of •NO2 with MbFeIIO2 and MbFeIIIOH2 was studied using the tert-butyl alcohol/NO3- system by way of reduction of nitrate by eaq- as well as through the dissociation of the t-BuOONO2 adduct (reaction -10). We first confirmed that neither t-BuOO‚, nor O2‚- nor the radiolytically produced H2O2 did react with MbFeIIO2 and MbFeIIIOH2 under pulse radiolysis conditions. Indeed, no changes in the absorbance at 582 and 602 nm were observed within 40 s after pulse-irradiation of a solution, which was saturated with 80% N2O and 20% O2 and contained 36 µM MbFeIIO2 or 30 µM MbFeIIIOH2, 40 mM tertbutyl alcohol and 12 mM PB at pH 7.1. In the case of MbFeIIIOH2, a slow increase of the absorption at 602 nm was observed at longer times due to a slow and progressive formation of ferrylmyoglobin in the reaction of MbFeIIIOH2 with the radiolytically produced H2O2.46-50 Oxymyoglobin reacts much slower with H2O2 than MbFeIIIOH2 51 and, therefore, its reaction with the radiolytically formed H2O2 could not be observed within the time scale of our pulse radiolysis experiment. Nevertheless, catalase was generally included in the reaction mixture to remove H2O2. Reaction of MbFeIIIOH2 with •NO2. The reaction of MbFeIIIOH2 with •NO2 was studied by pulse-irradiation of aerated solutions containing 30 µM MbFeIIIOH2, 20 mM NO3-, 40 mM tert-butyl alcohol and 12 mM PB at pH 7.1. No spectral changes were observed at 500-600 nm within 40 s after the pulse, indicating that •NO2 does not react with MbFeIIIOH2 under these experimental conditions. This conclusion was confirmed by steady-state irradiation of the same solution in the absence and the presence of 500 U/mL catalase. Only the appearance of the spectrum of MbFeIVdO was observed and this spectrum disappeared almost completely in the presence of catalase. Reaction of MbFeIIO2 with •NO2. Upon pulse-irradiation of aerated solutions containing 11-101 µM MbFeIIO2, 20 mM NO3-, 40 mM tert-butyl alcohol and 12 mM PB at pH 7.1, the (41) Mallard, W. G.; Ross, A. B.; Helman, W. P. NIST Standard References Database 40, Version 3.0, 1998. (42) Lymar, S. V.; Schwarz, H. A.; Czapski, G. J. Phys. Chem. A 2002, 106, 7245-7250. (43) Goldstein, S.; Lind, J.; Merenyi, G. J. Phys. Chem. 2004, 108, 17191725. (44) Logager, T.; Sehested, K. J. Phys. Chem. 1993, 97, 10047-10052. (45) Goldstein, S.; Czapski, G. Inorg. Chem. 1997, 36, 4156-4162. (46) King, N. K.; Winfield, M. E. J. Biol. Chem. 1963, 238, 1520-1528. (47) Yonetani, T.; Schleyer, H. J. Biol. Chem. 1967, 242, 1974-1979. (48) Krishna, M. C.; Samuni, A.; Taira, J.; Goldstein, S.; Mitchell, J. B.; Russo, A. J. Biol. Chem. 1996, 271, 26018-26025. (49) Onuoha, A. C.; Zu, X. L.; Rusling, J. F. J. Am. Chem. Soc. 1997, 119, 3979-3986. (50) Giulivi, C.; Cadenas, E. Free Radic. Biol. Med. 1998, 24, 269-279. (51) Yusa, K.; Shikama, K. Biochemistry 1987, 26, 6684-6688.

Kinetics and Mechanism of •NO2

ARTICLES

Figure 1. Reaction of MbFeIIO2 with •NO2. The Absorption changes at 602 nm observed upon pulse-irradiation (11.5 Gy) of aerated solution containing 27 µM [MbFeIIO2], 20 mM NO3-, 40 mM tert-butyl alcohol and 12 mM PB at pH 7.1. The optical path length was 3.1 cm.

Figure 2. Concentration-dependence of MbFeIIO2 reaction with •NO2. The observed first-order rate constant of the fast process as a function of [MbFeIIO2] as measured in aerated solutions containing 20 mM NO3-, 40 mM tert-butyl alcohol and 12 mM PB at pH 7.1 using 11.5 Gy/pulse.

absorption changes at 450-640 nm exhibited two consecutive first-order reactions, as demonstrated in Figure 1. The observed first-order rate constant of the fast process increased with increasing [MbFeIIO2]o (Figure 2). The second-order rate constant, which was obtained from the slope of the line in Figure 2, is (4.5 ( 0.3) × 107 M-1s-1. This value is considerably lower than k10 and k12 and, therefore, most •NO2 radicals react with t-BuOO‚ forming t-BuOONO2 rather than oxidizing MbFeIIO2. Hence, the extent of MbFeIIO2 oxidation observed during the fast process was relatively small. The slower first-order change of absorbance following the pulse had a rate constant kobs ) 0.15 ( 0.05 s-1 and was independent of [MbFeIIO2]. The major fraction of MbFeIIO2 oxidation took place during this stage. The •NO radicals reacting during this stage originated from a 2 relatively slow decomposition of t-BuOONO2 to yield t-BuOO• and •NO2.43 During this process the steady-state concentrations of both t-BuOO• and •NO2 radicals are considerably lower than that of MbFeIIO2. This implies that MbFeIIO2 will scavenge •NO quantitatively. Furthermore, since the rate of the reaction 2 of •NO2 with MbFeIIO2 is much larger than the rate of the reaction of •NO2 with t-BuOO‚, i.e., [MbFeIIO2] . [t-BuOO•], the rate-determining step of the slow process is expected to be characterized by k-10 ≈ 0.07 s-1.43 Despite this, the observed rate constant was ca. twice as high as this value. As will be explained in more detail below, this finding is best interpreted by assuming that every homolysis through reaction -10 gives rise to the consumption of 2 molecules of MbFeIIO2. The difference spectra, which were monitored at the end of the fast and the slow processes, are given in Figure 3A. Comparison of the transient difference spectra recorded respec-

Figure 3. A. Transient difference spectra monitored 2 ms and 40 s after the pulse of aerated solutions containing 29 µM MbFeIIO2, 20 mM NO3-, 40 mM t-butanol and 12 mM PB at pH 7.1 using 11.5 Gy/pulse and 3.1 cm optical path. B. Difference spectra at pH 7.1 between MbFeIIIOH2 and MbFeIIO2 (dashed) and MbFeIVdO and MbFeIIO2 (solid).

tively at 2 ms and 40 s after the pulse (Figure 3A) with the difference spectra between MbFeIIIOH2 and MbFeIIO2 or MbFeIVdO and MbFeIIO2 (Figure 3B) reveals that during the first 2 ms the reaction of MbFeIIO2 with •NO2 forms only MbFeIVdO. However, by 40 s after the pulse a mixture of MbFeIVdO and MbFeIIIOH2 is observed. This observation is further substantiated by a comparison between the ratios of the absorbance changes at specific wavelengths, such as ∆OD602/ ∆OD582, ∆OD630/∆OD582 and ∆OD602/∆OD563. To determine more accurately the yields of MbFeIVdO and MbFeIIIOH2 in the reaction mixture, the absorption spectrum of the irradiated solution was scanned by a diode array spectrophotometer 40 s after the pulse. Since NADH reduces MbFeIVdO to MbFeIIIOH2,48 the experiment was repeated with NADH being added to the solution 40 s after pulse-irradiation. The difference spectra monitored after pulse-irradiation of 29 µM MbFeIIO2, with or without 1 mM NADH being added after the irradiation, are presented in Figure 4. The spectra clearly demonstrate the reduction of the ferryl into the ferric state. The oxidation yield in the absence of NADH was calculated using the two isosbestic points of MbFeIVdO and MbFeIIIOH2 at 519 nm (∆519 ) 3 mM-1cm-1) and 616 nm (∆616 ) 2.6 mM-1cm-1), and was within the accuracy of the experimental measurements identical to the yield of MbFeIIIOH2 obtained upon the addition of NADH. The oxidation yield was independent of the dose and of [MbFeIIO2]o, when the latter varied between 29 and 94 µM. This yield was G ) 4.4 ( 0.4, which is about twice the yield of •NO2 reacting with MbFeIIO2. Using the known spectra of the various redox states at pH 7.1, we determined the ratio of the yield of MbFeIVdO to that of MbFeIIIOH2 by 40 s after the pulse as 1.06 ( 0.11. J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004 15697

Goldstein et al.

ARTICLES

Figure 4. Transient difference spectra monitored 40 s after the pulse (solid) and upon the addition of 1 mM NADH 40 s after the pulse (dashed). The aerated solutions contained 29 µM MbFeIIO2, 20 mM NO3-, 40 mM tertbutyl alcohol and 12 mM PB at pH 7.15. The dose was 11.5 Gy and the optical path length was 1 cm.

We, therefore, suggest that •NO2 adds to MbFeIIO2 to form MbFeIIIOONO2 as a short-lived intermediate (reaction 13), and k13 ) (4.5 ( 0.3) × 107 M-1s-1 was determined from the slope of the line in Figure 2. The intercept in this plot represents mainly the competition between MbFeIIO2 and t-BuOO• for •NO radicals. 2

MbFeIIO2 + •NO2 f MbFeIIIOONO2

(13)

Chemically, MbFeIIIOONO2 is akin to the short-lived peroxo adduct, which most probably is formed as an intermediate in the reaction of MbFeIIIOH2 with H2O2. It is widely believed that this adduct yields a ferryl having the radical site on the globin, •MbFeIVdO (eq 14).46-50 -H+,

-H2O

Figure 5. Concentration-dependence of MbFeIIIOH2 reaction with peroxynitrate. The observed first-order rate constant of the formation of the absorbance at 580 nm as measured upon pulse-irradiation of aerated solutions containing 40 mM NO3-, 0.1 M formate and 2 mM acetate buffer at pH 4.8 using 16 Gy/pulse.

during the fast process, k17 must be much smaller than k13. In conclusion, 104 M-1s-1 < k17 < 107 M-1s-1. The rate constant for the reaction of MbFeIIO2 with H2O2 is relatively small, namely k ) 20.8 M-1s-1 51, and considering the low concentration of the radiolytically formed H2O2, this reaction can be ignored during the spectrophotometric experiments. We note that the yield of MbFeIIIOH2 increased slowly with time at the expense of MbFeIVdO even in the presence of catalase. Within 1 h after the pulse it approached G(MbFeIIIOH2) ) 7.0, i.e., G(MbFeIIIOH2) ≈ 3 G(•NO2). We attribute this slow process to the comproportionation of MbFeIIO2 with MbFeIVdO (reaction 18).

MbFeIVdO + MbFeIIO2 + H2O + 2H+ f 2 MbFeIIIOH2 + O2 (18)

MbFeIIIOH2 + H2O2 98 -OH-

MbFeIIIOOH 98 MbFeVdO f •MbFeIVdO (14) On the basis of this analogy, we propose that MbFeIIIOONO2 decomposes via heterolysis of the O-O bond to form MbFeVdO and NO3- (reaction 15). Subsequently, MbFeVdO rapidly transforms into •MbFeIVdO (reaction 16), which oxidizes MbFeIIO2 (reaction 17).

MbFeIIIOONO2 f MbFeVdO + NO3- fast •

V

IV

MbFe dO f MbFe dO fast •

(15) (16)

MbFe dO + MbFe O2 + H2O f IV

II

MbFeIVdO + MbFeIIIOH2 + O2 (17) Hence, the redox species formed within 2 ms after the pulse should be •MbFeIVdO, which is expected to absorb similarly to MbFeIVdO at 450-700 nm.46,47 Then, within 40 s, reaction 17 produces equal amounts of MbFeIVdO and MbFeIIIOH2. The observed rate constant of the slow process was ca. twice as high as k-10 ) 0.07 s-1 and was almost unaffected by [MbFeIIO2]o ) 11-101 µM. The simplest way of explaining this finding is to assume that, as long as [MbFeIIO2]o > 11 µM, k17[MbFeIIO2]o . k-10 ) 0.07 s-1. With these assumptions 2 molecules of MbFeIIO2 will be consumed for every homolysis in reaction -10 and hence kobs ) 2k-10. According to this interpretation k17 must be significantly higher than 0.07/10-5 ≈ 104 M-1s-1. On the other hand, as no formation of MbFeIIIOH2 was observed 15698 J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004

To study reaction 18, we mixed 22 µM MbFeIVdO with 225 µM MbFeIIO2 at pH 7.1. The increase in the absorption at 633 nm followed first-order kinetics resulting in kobs ) (2.4 ( 0.6) × 10-3 s-1 and ∆OD∞ ) 0.16 ( 0.01, implying k18 ) 21.3 ( 5.3 M-1s-1. This value is similar to those determined for the comproportionation reactions of bovine and human oxyhemoglobin, namely 19 and 23 M-1s-1, respectively.52 The stoichiometry of Reaction 18 is obtained from the respective absorbance changes using ∆633 ) 3.4 mM-1cm-1 as ∆[MbFeIIIOH2]/ ∆[MbFeIVdO] ) 2.1 ( 0.1. Reaction of MbFeIIIOH2 with Peroxynitrate. O2NOOH is similar to HOOH in that both are relatively resistant to homolysis of the O-O bond. We, therefore, hypothesized that their reactions with MbFeIIIOH2 should also be similar. Peroxynitrate (O2NOO-/O2NOOH, pKa ) 5.9)44,45 was produced from superoxide (HO2•/O2•-, pKa ) 4.8) and •NO2 by pulse irradiation of aerated solutions containing 0.1 M formate, 40 mM NO3and 2 mM PB at pH 4.8-5.1 or 24 mM PB at pH 5.9-7.2. The reaction of MbFeIIIOH2 with peroxynitrate could be studied because its precursors, i.e., O2•- and •NO2, were found to react extremely slowly with MbFeIIIOH2. To study the reaction kinetics, the absorption changes at 580 nm upon pulse-irradiation of 22-105 µM MbFeIIIOH2 were followed. The rate of the build up of the absorption at 580 nm obeyed first-order kinetics, and kobs was linearly dependent on [MbFeIIIOH2]o (Figure 5). The (52) Giulivi, C.; Davies, K. J. Biol. Chem. 1990, 265, 19453-19460.

Kinetics and Mechanism of •NO2

ARTICLES

(Figure 6), whereas that of the slow decay process was independent of [MbFeIVdO]o with kobs ) 190 ( 20 s-1. These results were unaffected upon replacing 2 mM borate with 20 mM PB or by varying the pH from 9.7 to 10.3. In analogy with the suggestion of Herold et al.,54 we propose that MbFeIVdO is reduced by •NO2 via the formation of the transient species MbFeIIIONO2.

MbFeIVdO + •NO2 f MbFeIIIONO2 OH-

MbFeIIIONO2 98 MbFeIIIOH + NO3Figure 6. Concentration-dependence of MbFeIVdO reaction with •NO2. The observed first-order rate constant of the fast process as a function of [MbFeIVdO] as measured at 638 nm in N2O/O2 ([N2O] > 15 mM) saturated solutions containing 1 mM NO2-, and either 2 mM borate buffer at pH 9.7 or 20 mM PB at pH 10.3 using 16 Gy/pulse.

apparent bimolecular rate constant was unaffected by increasing the pH, and at pH > 6.5 the absorption yield was too small for accurate measurements. These observations demonstrate that O2NOOH and O2NOO- react with MbFeIIIOH2 at similar rates. At pH > 6.5 the absorption increase was relatively small due to an efficient competition of the self-decomposition of O2NOO(k ) 1.3 s-1 25 °C)44,45 with its reaction with MbFeIIIOH2. The species formed after the pulse was identified as MbFeIVdO. The yield of MbFeIVdO exceeded 95% of the peroxynitrate initially formed (∆580 ) 5.4 mM-1cm-1 at pH 4.8 between the extinction coefficients of MbFeIVdO and MbFeIIIOH2). We, therefore, suggest that MbFeIVdO is formed by Reaction 19 followed by reactions 15 and 16.

MbFeIIIOH2 + O2NOOH f MbFeIIIOONO2 + H2O + H+ (19) From the slope of the line in Figure 5, we obtained k19 ) (4.6 ( 0.3) × 104 M-1s-1. This value is close to 1.03 × 104 M-1s-1 11 and (6.6 ( 0.3) × 104 M-1s-1,22 the rate constant of MbFeIIIOH2 reacting with ONOOH/ONOO-, which are similar in size and acid base properties to O2NOOH/O2NOO-. As discussed above, MbFeIIIOONO2 undergoes rapid heterolysis of the O-O bond to form •MbFeIVdO (Reactions 15-16). Even in the absence of potential reductants, such as MbFeIIO2, the half-life of •MbFeIVdO is relatively short,47,50,53 eventually yielding the more stable nonradical ferryl species MbFeIVdO. Reaction of MbFeIVdO with •NO2. Since MbFeIVdO was found to be stable in the presence of 1 mM nitrite at pH > 9.7 for the duration of the experiment, •NO2 radicals could be generated through the oxidation of nitrite by •OH. Relatively small volumes of aerated solutions of MbFeIVdO were added to N2O-saturated solutions containing 1 mM nitrite and 2 mM borate at pH 9.7 or 20 mM PB at pH 10.3, which were pulseirradiated within a few minutes of preparation. In these solutions the concentration of N2O was higher than 16 mM whereas that of O2 did not exceed 0.4 mM. A rapid first-order build-up of a transient absorption was observed at 638 nm, which decayed via a first-order reaction leaving behind a residual persistent absorption. The observed first-order rate constant of the fast formation process increased with increasing [MbFeIVdO]o (53) Gunther, M. R.; Sturgeon, B. E.; Mason, R. P. Free Radic. Biol. Med. 2000, 28, 709-719.

(20) (21)

Hence, k21 ) 190 ( 20 s-1 and from the slope of the line in Figure 6 we determined k20 ) (1.2 ( 0.2) × 107 M-1s-1. The intercept in this plot represents the competition between the reaction of •NO2 with MbFeIVdO (reaction 20) and the dimerization/hydrolysis of •NO2 (reaction 22).

2 •NO2 + H2O f NO2- + NO3- + 2H+ 2k22 ) 1.3 × 108 M-1s-1 41 (22) Reactions 20 and 21 imply a stoichiometry of [MbFeIIIOH]/ [•NO2] ) 1. We determined the stoichiometry of this process using γ-radiolysis. Under such conditions, the dimerization/ hydrolysis of •NO2 (reactions 22) and the reaction of •NO2 with •NO (Reactions 23 and 24), which is formed via reaction 5, can be ignored. •

NO + •NO2 h N2O3 k23 ) 1.1 × 109 M-1s-1 k-23 ) 8.4 × 104 s-1 55 (23)

N2O3 + OH- f 2NO2- + H+ k24 ) 2 × 103 + 1 × 108[OH-] s-1 56 (24) Upon γ-radiolysis of a solution saturated with 80% N2O and 20% O2 containing 43.2 µM MbFeIVdO, 1.2 mM NO2-, 1000 U/mL catalase and 5 mM borate buffer at pH 9.8, the yield of MbFeIIIOH increased linearly with the dose. The difference spectra obtained following irradiation were compared to that measured for 43.2 µM MbFeIVdO under the same experimental conditions. From the slope of the lines in Figure 7 we determined G(MbFeIIIOH) ) 6.7 ( 0.3. This high value implies that G(MbFeIIIOH) ≈ GOH + Ge + GH. As the H• atoms in this system convert NO2- to •NO (Reaction 5), the measured G-value reflects the ability of •NO to reduce additional MbFeIVdO to MbFeIIIOH. A similar experiment performed using a solution containing 21.6 µM MbFeIVdO at pH 10.25 (5.7 mM borate, 1.3 mM NO2-, 1000 U/mL catalase, 88% N2O, 12% O2) resulted in G(MbFeIIIOH) ) 5.9 ( 0.2. While still very high, this value is somewhat lower than the one obtained at higher [MbFeIVdO]o, and could be attributed to a higher scavenging efficiency of •NO and •NO by 43.2 µM MbFeIVdO. 2 Herold et al.54 studied by pulse radiolysis the reaction of MbFeIVdO with •NO2 in N2O-saturated solutions containing 1.5 µM MbFeIVdO, 10 mM nitrite and 50 mM borate at pH 9.5. They assumed that, practically all the primary radicals formed by the pulse, produced exclusively •NO2 (150 µM). In (54) Herold, S.; Exner, M.; Nauser, T. Biochemistry 2001, 40, 3385-3395. (55) Gratzel, M.; Taniguch. S.; Henglein, A. Berichte Der Bunsen-Gesellschaft Fur Physikalische Chemie 1970, 74, 488-492. (56) Treinin, A.; Hayon, E. J. Am. Chem. Soc. 1970, 92, 5821-&. J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004 15699

Goldstein et al.

ARTICLES

Figure 7. Reaction of MbFeIVdO with •NO2. The difference spectra obtained upon γ-irradiation of a solution saturated with 80% N2O and 20% O2 containing 43.2 µM MbFeIVdO, 1.2 mM NO2-, 500 U/mL catalase and 5 mM borate at pH 9.8 for 1, 2, 3 and 4 min at dose rate of 9.4 Gy min-1. The bold line is the difference spectrum between 43.2 µM MbFeIIIOH and 43.2 µM MbFeIVdO under the same conditions. The dependence of the yield of MbFeIIIOH on the dose resulted in G(MbFeIIIOH) ) 6.7 ( 0.3.

this single experiment they obtained k20 ) 1 × 107 M-1s-1, which is close to our value. However, their derived value of k21 ) 35 ( 5 s-1 is seen to be significantly lower than that determined in the present work, i.e., k21 ) 190 ( 20 s-1. We note, however, that under their experimental conditions ca. 15% of eaq- and all H• radicals react with NO2- to form •NO. Hence, the initial concentrations of •NO2 and •NO should be about 140 and 26 µM, respectively. Therefore, the hydrolysis of •NO2 (Reaction 22) and the fast reaction of •NO with •NO2 (Reaction 23) have to be taken into account. Under these circumstances, all •NO2 decayed within 2 ms, and Herold et al.54 could not have determined k20 by assuming pseudo-first-order conditions with [•NO2]o ) 150 µM. Most importantly, due to the poorly defined radical mixture their derived value for k21 is bound to be seriously distorted. By contrast, given that in our measurements reaction 20 proceeds in clean pseudo-first-order conditions and reactions 20 and 21 are well-separated, the present value of k21 ) 190 ( 20 s-1 is believed to be rather accurate. In another set of experiments Herold et al.54 reacted MbFeIIO2 with •NO and observed an intermediate, which decayed with k ) 205 ( 5 s-1. Since this intermediate appeared to decay much faster than the one obtained in reaction 20, Herold et al.54 argued that the former must be MbFeIIIOONO. However, given that in the present study the two rate constants are essentially identical, we suggest that in both situations the only observable intermediate is MbFeIIIONO2. We note that the intermediate that Herold et al.54 produced in the reaction between MbFeIIO2 with •NO decayed much faster at low than at high pH. In fact, this intermediate was not even seen at pH 7. Similar pH dependences were reported for the decay of other anion adducts to MbFeIII, 15700 J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004

such as MbFeIIINO240 or MbFeIIIN3.57 Upon raising the pH, deprotonation at various sites occurs, which changes the overall charge of the protein. As the heterolysis of the adduct involves charge separation, this process is expected to be sensitive to the overall charge, which accounts for these observations. On the other hand, homolysis of the putative MbFeIIIOONO intermediate would not entail charge separation. Consequently, no obvious rationalization accounts for its apparent pHdependence. Our assignment of MbFeIIIONO2 to the observable intermediate implies that MbFeIIIOONO must undergo completely or at least to a major extent an instantaneous conversion to its isomer, MbFeIIIONO2, which has a much lower free energy than MbFeIIIOONO. This conversion is believed to be initiated by homolysis of the O-O bond to yield a geminate pair of MbFe(IV)dO + •NO2. Between 80%11 and 100%54 of the latter collapses to the cage product MbFeIIIONO2, while the rest, if any, escapes out of the solvent cage as freely diffusible MbFe(IV)dO and •NO2. We estimate that the O-O bond in MbFeIIIOONO is significantly weaker than even in alkylperoxynitrites. In fact, while the homolysis of alkyl peroxynitrites is still slightly endergonic, the corresponding homolysis of MbFeIIIOONO is almost certainly exergonic. These conclusions are borne out by the following arguments. The equilibrium constant for reaction 25 at pH 7.0 and 25 °C has been determined to be 974 atm-1.58

MbFeII + O2 h MbFeIIO2 K25 ) 974 atm-1

(25)

Using reduction potentials at pH 7 vs. NHE with E°′(MbFeIIIOH2/ MbFeII) ≈ 0.0 V,59 E°′(MbFeIVdO/MbFeIIIOH2) ≈ 0.90 V,60 and E°′(O2/H2O) ) 0.815 V,61 we calculate E°′(26) ≈ 1.09 V.

MbFeIIO2 + 2e- + 2H+ h MbFeIVdO + H2O E°′ ≈ 1.09 V (26) Similarly, from ∆Gf°(•NO2, aq) ) 15.1 kcal/mol,62 ∆Gf°(•NO, aq) ) 24.4 kcal/mol62 and ∆Gf°(H2O, l) ) -56.7 kcal/mol,61 we calculate E°′(27) ) 0.61 V. •

NO2 + 2e- + 2H+ h •NO + H2O E°′ ) 0.61 V (27)

Combining E°′(26) with E°′(27) we obtain ∆G° ≈ -22 kcal/ mol for the overall oxygen transfer reaction 28

MbFeIIO2 + •NO h MbFeIVdO + •NO2

(28)

Reaction 28 can be subdivided into reactions 29 and 30:

MbFeIIO2 + •NO h MbFeIIIOONO

(29)

MbFeIIIOONO h MbFeIVdO + •NO2

(30)

We can generalize reactions 29 and 30, where we shall treat (57) Lin, J.; Merryweather, J.; Vitello, L. B.; Erman, J. E. Arch. Biochem. Biophys. 1999, 361, 148-158. (58) Schenkman, K. A.; Marble, D. R.; Burns, D. H.; Feigl, E. O. J. Appl. Physiol. 1997, 82, 86-92. (59) Blankman, J. I.; Shahzad, N.; Dangi, B.; Miller, C. J.; Guiles, R. D. Biochemistry 2000, 39, 14799-14805. (60) He, B.; Sinclair, R.; Copeland, B. R.; Makino, R.; Powers, L. S.; Yamazaki, I. Biochemistry 1996, 35, 2413-2420. (61) Dean, J. A., Ed. Lange’s Handbook of Chemistry, 13 ed.; McGraw-Hill: New York. (62) Stanbury, D. M. AdV. Inorg. Chem. 1989, 33, 69-138.

Kinetics and Mechanism of •NO2

ARTICLES

the cases for X ) H, alkyl or MbFe

Scheme 1

XO2 + •NO h XOONO

(31)

XOONO h XO + •NO2

(32)

From published data43,62-64 we calculate ∆G°(31) ) -18.3 and ca. -17.3 kcal/mol for X ) H and X ) alkyl, respectively. The corresponding values for ∆G°(32) come out as 14.1 and ca.. 4.3 kcal/mol. We note that, upon replacing an alkyl group by MbFe, the exergonicity of the overall oxygen transfer reaction increases by 9 kcal/mol, i.e, ∆G° goes from - 13 to -22 kcal/ mol. Furthermore, from a comparison of H to alkyl, the homolysis reaction 32 would appear substantially more sensitive to X than reaction 31. Since in the former reaction the integrity of the O-O bond is disrupted, while in the latter it is not, this finding is not unexpected. We have no experimental data for ∆G°(29) and ∆G°(30). However, the rate of O-N homolysis of CraqOONO22+ into CraqOO2+ and •NO2 is reported to be faster by ca. 3 orders of magnitude, i.e., ∼180 M-1s-1,65 than that for ROONO2 into ROO• and •NO2, i.e., ∼0.1 M-1s-1.43 This shows that the N-O bond in transition metal peroxynitrates is probably weaker than in alkyl peroxynitrates. It is reasonable to assume that the same applies for the corresponding peroxynitrites. Hence, we are confident that ∆G°(29) > -17 kcal/ mol. Consequently, ∆G°(30) should be less than -5 kcal/mol, and on this basis the homolysis reaction 30 is predicted to be significantly exergonic. Given that the lifetime of alkyl peroxynitrites is below 1 µs,43,63 the lifetime of MbFeIIIOONO is expected to be even shorter. Hence, this intermediate should be undetectable under any experimental circumstances. The above reasoning concerns the production of free XO and •NO . However, the rate-determining step in the decay of 2 XOONO is homolysis to yield a geminate pair. It is seen that, at a given exergonicity of reaction 32 (or 30), the smaller the yield of cage escape the larger the exergonicity of homolysis to form the geminate pair. For X ) alkyl, the yield of cage escape is ca. 15%,43,63 while for X ) MbFe it is 0-20%.11,66 For the highest yield the relative exergonicities of Reaction 32 and geminate pair formation are about the same for X ) alkyl and X ) MbFe. Should the escape yield for X ) MbFe turn out to be very small, perhaps below 1%, the exergonicity of its homolysis to form the geminate pair increases even further relative to X ) alkyl. Thus, the conclusion about the short lifetime of MbFeOONO remains even more valid. Finally, we note that the exergonicity of homolysis precludes the reverse reaction -30 from taking place at a significant rate. This is consistent with the exclusive formation of MbFeIIIONO2 in reaction 20. Conclusions

The rate constant for reaction of •NO2 with oxymyoglobin was determined to be k13 ) (4.5 ( 0.3) × 107 M-1s-1, and is among the highest rate constants measured for •NO2 with any biomolecule, i.e., similar to those for the reaction of •NO2 with glutathione, cysteine, and urate at pH 7.4.67 As proposed in Scheme 1, the reaction of •NO2 with oxymyoglobin forms MbFeIIIOONO2, which undergoes rapid heterolysis along the O-O bond to yield MbFeVdO and NO3-. The perferrylmyoglobin rapidly transforms into •MbFeIVdO, which oxidizes another oxymyoglobin (104 M-1s-1 < k17 < 107 M-1s-1) and generates equal amounts of ferrylmyoglobin and metmyoglobin. Over a much longer time period, the ferrylmyoglobin disappears by way of a relatively slow comproportionation with oxymyoglobin (k18 ) 21.3 ( 5.3 M-1s-1). Eventually, each •NO2 radical converts three oxymyoglobin molecules into metmyoglobin (Scheme 1). The same adduct, namely MbFeIIIOONO2, is formed via the reaction of metmyoglobin with O2NOOH (k19 ) (4.6 ( 0.3) × 104 M-1s-1), and this reaction generates ferrylmyoglobin and nitrate (Scheme 1). The reaction of •NO2 with ferrylmyoglobin (k20 ) (1.2 ( 0.2) × 107 M-1s-1) yields MbFeIIIONO2, which in turn dissociates (k21 ) 190 ( 20 s-1) into metmyoglobin and NO3-. The same intermediate is produced in the reaction of MbFeIIO2 with •NO. The potential physiological implications of this chemistry are that hemeproteins can be instrumental in detoxifying •NO2, and hence in pre-empting deleterious processes, such as nitration of proteins. This conclusion is substantiated by the observation that the reactions of •NO2 with the various oxidation states of myoglobin lead to the formation of metmyoglobin, which, though not functional in the gas transport, is nevertheless nontoxic at physiological pH. Acknowledgment. This research was supported by grants from the U.S.-Israel Binational Science Foundation (No. 2002013). G.M. is grateful for the financial support given by the Swedish Science Council (VR).

The present study focused on the stoichiometry, kinetics and mechanisms of •NO2 reacting with oxymyoglobin and ferrylmyoglobin. The interconversion among the various oxidation states of myoglobin that is prompted by nitrogen oxide species is remarkable (Scheme 1).

Note Added after ASAP Posting. After this paper was posted ASAP on November 11, 2004, two authors’ affiliations were corrected, the label for eq 21 was added, and the E°′ notation in the text below eq 25 was corrected. The corrected version was posted November 16, 2004.

(63) Merenyi, G.; Lind, J.; Goldstein, S. J. Am. Chem. Soc. 2002, 124, 40-48. (64) Merenyi, G.; Lind, J.; Czapski, G.; Goldstein, S. Inorg. Chem. 2003, 42, 3796-3800. (65) Pestovsky, O.; Bakac, A. Inorg. Chem. 2003, 42, 1744-1750. (66) Herold, S. Comptes Rendus Biologies 2003, 326, 533-541.

JA046186+ (67) Ford, E.; Hughes, M. N.; Wardman, P. Free Radic. Biol. Med. 2002, 32, 1314-1323. J. AM. CHEM. SOC.

9

VOL. 126, NO. 48, 2004 15701